Tests to compare aws-lambda supported languages (micronaut, nodejs, python) for memory use and performance

Use cases to be tested

• /echo/{text} - baseline
• /find-primes-below/{number} - computational work
• /food-trivia-random/{apiKey} - goto spoonacular.com to get an apiKey I'm not affiliated with these folks in anyway, just needed a simple api to test http client

Technologies Used:

• AWS Lambda - serverless functions https://docs.aws.amazon.com/lambda/index.html
• GraalVm v20.1.0 - compile to native code https://www.graalvm.org/docs/release-notes/20_0/
• Docker (installation required) - needed to run graalvm in amazonlinux environment
  • Toolbox - https://github.com/docker/toolbox/releases
  • Desktop - https://docs.docker.com/release-notes/
• VSCode (optional) - 2 very useful extensions: Docker (view container, images) and AWS Explorer (view/invoke lambda functions)
  • Download - https://code.visualstudio.com/download
• AWS SAMCLI (optional) - test lambda functions locally
  • Installation - https://docs.aws.amazon.com/serverless-application-model/latest/developerguide/serverless-sam-cli-install.html
• AWS CLI - used to upload lambda functions to AWS
  • Installation (required) - https://docs.aws.amazon.com/cli/latest/userguide/install-cliv2.html
• Gradle - build tool for micronaut, nodejs and python

Other Installation:

• Micronaut v2.0.0 - https://micronaut.io/download.html
  • OpenJDK 14 - https://developers.redhat.com/products/openjdk/download
• nodejs v12.16.3 - https://nodejs.org/en/download/
• python v3.8 - this is installed w/ conda, or Visual Studio
  • Anaconda Install - https://docs.anaconda.com/anaconda/install/

Installation Instructions

The only installation that wasn't straight forward, was Docker. So good luck ... kidding. I'll help as much as I can. I'm using Windows Home, so I chose Docker Toolbox (versus Docker Desktop).

Docker Toolbox Installation Notes

• Download Docker Toolbox (for Windows, Mac or Linux) - https://docs.docker.com/toolbox/
• Check make sure your cpu has virtualization enabled -- THIS IS VERY IMPORTANT
• After installation, check to be sure Docker is working
  • Run Oracle VM VirtualBox (Console App)
  • Run Docker Quickstart Terminal -- wait for VirtualBox to finish (takes a minute or two)
  • In the docker terminal: $docker run hello-world -
    • if problems, reset the default machine in the Console App OR run the following commands in the terminal
      • $ docker-machine rm default
      • $ docker-machine create --driver virtualbox default
    • Check VS Code (install Docker extension), you should see the docker containers, images, etc
• Update the VM settings for the Virtual Box that you're using (default)
  • Open Oracle VM Virtual Box
  • Close (if running) your VM - I use the default
    • on windows, right click on VM (default) and select "Close" (power off)
  • Click settings
    • System tab -
      • increase memory to at least 4Gb - if you are unable to change settings, it's because your VM is still running
      • increase cpu to at least 4 (or half system total)
    • Display tab - increase your video memory (I'm using 32mb - anything more than 8mb)
    • Storage tab - I've spent many hours trying to increase disk.vmdk - the boot2docker doesn't like it (won't let you connect)
    • Shared folders - mount any drives that you need. Docker needs access to the paths for java, micronaut, etc

I found this thread to be very helpful: docker/for-win#204

Handy Docker commands

$ docker-machine ip --- you'll need this if you want to run the app in docker, and test in the browser http://machineip:8088/

Micronaut Build

I'm planning on update the build.gradle files so that you can manage all the builds from the root.
I haven't done it yet, so for now, you'll have to build in the project folders.

• Run the Oracle VM - start the VM (default) - I start it headless
• Run the Docker Quickstart Terminal
• Wait for it to load ...
• cd to aws-lamda-graalvm/micronaut folder
• for brand new installation

./gradlew assemble (docker env)
./docker-build.sh  (docker env)
mkdir build

After code changes, I like to test/debug locally (local env)

./gradlew test  

Build and create a folder.zip and upload to Amazon (docker env)

$./docker-build.sh
$docker run --rm --entrypoint cat micronaut  /home/application/function.zip > build/function.zip
$aws lambda create-function \
  --function-name micronaut \
  --zip-file fileb://build/function.zip \
  --handler function.handler \
  --runtime provided \
  --role [enter your aws-lambda-role]

For updates:

$aws lambda update-function-code \
  --function-name micronaut \
  --zip-file fileb://build/function.zip

There's also a deploy.sh script that does the above

Use the following to invoke the lammbda function

aws lambda invoke --function-name micronaut --payload ''

Use the following payloads:

{
  "resource": "/echo/blah",
  "path": "/echo/blah",
  "httpMethod": "GET"
}

{
  "resource": "/find-primes-below/999",
  "path": "/find-primes-below/999",
  "httpMethod": "GET"
}

{
  "resource": "/food-trivia-random/{enter your apiKey}",
  "path": "/food-trivia-random/{enter your apiKey}",
  "httpMethod": "GET"
}

I'll update the doc once I wire this to AWS API-Gateway (REST API) and/or AWS AppSync (GraphQL)

I'll update when I add the nodejs and python tests